Malaysia SEACO-CH20 Smartwatch Feasibility Study
First thing to do is to get access to the SEACO RDSF - ask the PI for this.
Then you'll need to mount it somewhere - the mount location that I used is in userconf.yml.
Run conda create -f environment.yml to create an environment that contains all the required packages,
including R, python and the required libraries.
If you have trouble with this, try installing R and python separately.
Approximately correspond to the order of things in the paper:
- demographic_summary.ipynb: stats on the participant demographics
- survey.ipynb: quantitative results from the survey
- meal_stats.ipynb: statistics on the numbers of meals, snacks, etc. per day
- three_level_model.ipynb: linear models for response rate as the study progesses
If you're just interested in the linear model: three_level_models.R.
The notebook three_level_model.ipynb creates a .csv file for boolean response/not for each hour/day/participant.
It also prints some stuff like the median of the mean response rate per day, etc. (which are also in the paper).
Then run Rscript analysis_utils/r/three_level_models.R to get the full output.
Unfortunately the code here is a bit of a mess so you might have to do a fair amount of
reading through source code - there's a helper library in analysis_utils/, somewhere
There are also other notebooks for things like the survey responses, the participant demographics (age, sex etc). If you're reading this far because you've inherited the project and have to answer reviewer comments, good luck
There are also additional notebooks and R scripts in old_stuff/, but these are either outdated or irrelevant for the paper.
I've kept them here just in case...
There are two config files:
config.yml- containing configuration (filepaths etc.) that you won't need to changeuserconf.yml- containing configuration that you might need to change, depending on e.g. where you have mounted the SEACO CH-20 RDSF